Validation of Geometric Mistuning Reduced-Order Models for Single and Dual Flow-Path Integrally Bladed Rotors

Abstract

This efforts validates the utilization of tuned secondary modal reductions in building approximate, highly reduced-order models of geometrically mistuned single and dual flowpath integrally bladed rotors (IBRs). The blades of each IBR are geometrically mistuned according to physical measurements taken from industrial hardware. A probabilistic study is conducted on a population of single and dual flow-path IBRs with random blade geometry perturbations; the mistuned forced response levels are calculated from different fidelity models, including: full finite element models, lower fidelity Craig-Bampton component mode synthesis models, and the lowest fidelity models reduced with tuned-secondary modal reductions. Statistical comparisons of the mistuned forced response distributions are carried out to show any differences in the lowest fidelity models’ ability to capture the forced response levels of a fleet of single and dual flow-path IBRs. Results indicate the lowest fidelity, highly-reduced models provide accurate forced response predictions for a population of rotors. Lastly, correlations between mistuned response prediction error for the single flow-path IBR and error metrics composed of secondary modes used in the lowfidelity models are investigated to determine if there are linkages between model accuracies and secondary modes used in model creation.